I think that it's always good to use caution when working on a research project that involves people as test subjects. Whatever you do, you need to make sure that you an objective, accurate way of measuring results and gathering data. For instance, a survey is very subjective and you have no way of really gauging if the people are learning better (they could be having a bad day, they could have not slept the night before, they could be hungry or stressed about some event coming up, etc.) Additionally with people, you need to run multiple trials because with humans, there are so many variables. From all these people, you need consent to use their results if you intend on using them in a lab-write up.

Secondly, I'm curious as to what kind of experiment you actually intend to run. Just have them listen to music and take a survey? I assume that your hypothesis would be that listening to music improves peoples' memory and learning? What is going to be your control group?

I think you have an interesting idea. I just don't know if it it's feasible to run as an actual research project or if there are too many variables involved.